An Integrated Approach for Automatic Construction of Bilingual Chinese-English WordNet

摘要

This paper compares various approaches for constructing Chinese-English bilingual WordNet. First, we implement three independent approaches that translate English WordNet to Chinese WordNet automatically, including Minimum Distance (MDA), Intersection (IA) and Words Co-occurrence (WCA). Minimum Distance compares the gloss of synset with the explanations of words from dictionaries. Intersection chooses the intersection part of Chinese in a synset. Words Co-occurrence counts the results of Chinese and English words from Google. Then, we integrate these three approaches into an integrated one, which is named MIWA. Experimental results show that the integrated approach MIWA has better performance: F-measure reaches 0.615, which is higher than that of each independent one.